Comparative Analysis of MobileNetV3 and EfficientNetv2B0 in BISINDO Hand Sign Recognition Using MediaPipe Landmarks
DOI:
https://doi.org/10.30871/jaic.v10i1.11878Keywords:
Sign Language Recognition, EfficientNetV2B0, MobileNetV3, MediaPipe, Deep LearningAbstract
Sign language is a vital communication medium for individuals with hearing and speech impairments. In Indonesia, more than 2.6 million people experience hearing disabilities, most of whom rely on Bahasa Isyarat Indonesia BISINDO for daily interaction. However, limited public understanding and the scarcity of professional interpreters continue to hinder inclusive communication. Recent advancements in computer vision and deep learning have enabled camera-based sign language recognition systems that are more affordable and practical compared to sensor-glove solutions. this study presents a comparative analysis between EfficientNetV2-B0 and MobileNetV3-Large in recognizing BISINDO hand sign alphabets using MediaPipe landmarks. The dataset was derived from BISINDO video recordings, from which hand landmarks were extracted using MediaPipe Hands and subsequently converted into two-dimensional skeletal images. In total, 10,309 skeletal images representing BISINDO alphabets A–Z were generated and used for model training and evaluation. Both models were trained under identical configurations using TensorFlow. The results show that MobileNetV3-Large achieved 89.67% test accuracy and an F1-score of 89.76%, while EfficientNetV2-B0 obtains 95.98% test accuracy and an F1-score of 95.93%. These findings highlight the trade-off between the higher classification accuracy of EfficientNetV2-B0 and the superior computational efficiency of MobileNetV3-Large. This research contributes to the development of lightweight, high-performance BISINDO recognition systems for assistive communication applications.
Downloads
References
[1] R. Fahlevi and C. Rozikin, “Identifikasi isyarat tangan bisindo dengan algoritma cnn dan transfer learning menggunakan mobilenetv2,” JATI (Jurnal Mahasiswa Teknik Informatika), vol. 9, no. 4, pp. 6592–6597, May 2025, doi: 10.36040/JATI.V9I4.14095.
[2] A. Saleh, “A Comparative Analysis of CNN and SVM for Static Sign Language Recognition Using MediaPipe Landmarks,” Journal of Intelligent System and Telecommunication, vol. 1, no. 2, pp. 225–238, Jun. 2025, doi: 10.26740/JISTEL.V1N2.P225-238.
[3] R. Sutjiadi, “Android-Based Application for Real-Time Indonesian Sign Language Recognition Using Convolutional Neural Network,” TEM Journal, vol. 12, no. 3, pp. 1541–1549, Aug. 2023, doi: 10.18421/TEM123-35.
[4] B. Sundar and T. Bagyammal, “American Sign Language Recognition for Alphabets Using MediaPipe and LSTM,” Procedia Comput Sci, vol. 215, pp. 642–651, Jan. 2022, doi: 10.1016/J.PROCS.2022.12.066.
[5] J. Bora, S. Dehingia, A. Boruah, A. A. Chetia, and D. Gogoi, “Real-time Assamese Sign Language Recognition using MediaPipe and Deep Learning,” Procedia Comput Sci, vol. 218, pp. 1384–1393, Jan. 2023, doi: 10.1016/J.PROCS.2023.01.117.
[6] O. Yusuf, M. Habib, and M. Moustafa, “Real-Time Hand Gesture Recognition: Integrating Skeleton-Based Data Fusion and Multi-Stream CNN,” Oct. 2024, Accessed: Oct. 26, 2025. [Online]. Available: https://arxiv.org/pdf/2406.15003v2
[7] D. Joan, V. Vincent, K. J. Daniel, S. Achmad, and R. Sutoyo, “BISINDO Hand-Sign Detection Using Transfer Learning,” 8th International Conference on Recent Advances and Innovations in Engineering: Empowering Computing, Analytics, and Engineering Through Digital Innovation, ICRAIE 2023, pp. 1–7, Dec. 2023, doi: 10.1109/ICRAIE59459.2023.10468194.
[8] T. Shahriar, “Comparative Analysis of Lightweight Deep Learning Models for Memory-Constrained Devices,” May 2025, Accessed: Oct. 26, 2025. [Online]. Available: https://arxiv.org/pdf/2505.03303
[9] M. Al-Hammadi et al., “Deep Learning-Based Approach for Sign Language Gesture Recognition With Efficient Hand Gesture Representation,” IEEE Access, vol. 8, pp. 192527–192542, Oct. 2020, doi: 10.1109/ACCESS.2020.3032140.
[10] M. Tan and Q. V. Le, “EfficientNetV2: Smaller Models and Faster Training,” Proc Mach Learn Res, vol. 139, pp. 10096–10106, Apr. 2021, Accessed: Oct. 26, 2025. [Online]. Available: https://arxiv.org/pdf/2104.00298
[11] J. Shin, A. S. M. Miah, M. H. Kabir, M. A. Rahim, and A. Al Shiam, “A Methodological and Structural Review of Hand Gesture Recognition Across Diverse Data Modalities,” IEEE Access, vol. 12, pp. 142606–142639, 2024, doi: 10.1109/ACCESS.2024.3456436.
[12] S. Sharma and S. Singh, “ISL recognition system using integrated mobile-net and transfer learning method,” Expert Syst Appl, vol. 221, pp. 119772–119772, Mar. 2023, doi: 10.1016/J.ESWA.2023.119772.
[13] I. Rizka Fadhillah, M. Muharrom Al Haromainy, and H. Maulana, “Implementasi model transfer learning efficientnet untuk pendeteksian bahasa isyarat indonesia (bisindo) pada perangkat android,” JATI (Jurnal Mahasiswa Teknik Informatika), vol. 8, no. 4, pp. 7816–7822, Aug. 2024, doi: 10.36040/JATI.V8I4.10463.
[14] R. A. Lashaki, Z. Raeisi, N. Razavi, M. Goodarzi, and H. Najafzadeh, “Optimized classification of dental implants using convolutional neural networks and pre-trained models with preprocessed data,” BMC Oral Health, vol. 25, no. 1, pp. 1–22, Dec. 2025, doi: 10.1186/S12903-025-05704-0/TABLES/3.
[15] M. K. Habib, O. Yusuf, and M. Moustafa, “Skeleton-Based Real-Time Hand Gesture Recognition Using Data Fusion and Ensemble Multi-Stream CNN Architecture,” Technologies 2025, Vol. 13, Page 484, vol. 13, no. 11, p. 484, Oct. 2025, doi: 10.3390/TECHNOLOGIES13110484.
[16] M. Pu, C. Y. Chong, and M. K. Lim, “Robustness Evaluation in Hand Pose Estimation Models using Metamorphic Testing,” Mar. 2023, Accessed: Jan. 09, 2026. [Online]. Available: http://arxiv.org/abs/2303.04566
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Alief Khairul Fadzli, Majid Rahardi
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) ) that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
